The Rosenhan experiment is a study conducted in 1973 in order to examine the validity of psychiatric diagnosis. Researcher David Rosenhan suspected that psychiatric criteria applied in diagnosis are not as objective as believed but rather context and personally determined.
In order to test his hypothesis Rosenhan wanted to commit healthy people to mental institutions and see how they are treated (hence the title of his article: "On Being Sane in Insane Places"). His study included two experiments testing unknowing real hospitals. In his first experiment Rosenhan sent healthy people to fake psychiatric symptoms which disappeared after they were hospitalized. These fake patients continued to be treated as abnormal though they presented no symptoms and were in fact completely sane (for a detailed account of the experiment see our summary of "On Being Sane in Insane Places" part 1). After the results of the first experiment, the second Rosenhan experiment informed institutions that fake patients will come to them within a given time frame. Rosenhan asked the institutes to try and pick up on the impostors and though a relatively high number of patients were suspected to be sane, Rosenhan in fact never sent anyone (for a detailed account of the experiment see our summary of "On Being Sane in Insane Places" part 2).
What the Rosenhan experiment shows is not only the unreliability of mental diagnosis (and maybe medical diagnosis in general), but something much deeper. In his article Rosenhan talks about the problem of "psychiatric labeling", a phenomenon by which once someone is considered to be insane the way he is perceived and treated by his surroundings will work to fulfill that prophecy. In the most broader sense, the Rosenhan study demonstrates how once we have an initial verdict on someone or something, everything we see from that point on will be understood based on that initial judgment, be it misguided or not. Rosenhan's experiment pushed forward the anti-psychiatry movement and the discourse on psychiatric labeling, but its meaning and implications run much deeper.